PathFinder AI: An Agentic AI Career Guide and Technical Interview Assistant

The current job market environment is one that lacks cohesion among career resources, generalist tips, and lack of personalized, efficient career preparation solutions. Traditional job websites and inflexible resume screeners neglect to consider time as a crucial variable, leaving individuals with disorganized learning material and without any clear roadmap to prepare for interviews. In this paper, we present PathFinder AI – an innovative career platform based on Agentic AI technology and intended to provide affordable, personal guidance. PathFinder AI consists of a three-layer architecture system, using the Node.js and Express frameworks combined with large language models through APIs offered by OpenAI and Gemini. An important contribution of this project is the Deadline Aware Scheduling Algorithm, which reverse-engineers a preparation curriculum anchored to the user's specific interview date. Through semantic skill-gap analysis — comparing vectorized user resumes against target job descriptions — the system pinpoints high-priority competencies and assigns curated learning resources in real time. The platform further incorporates five core agentic modules: a voice-driven AI Mock Interviewer built on the Whisper Medium speech recognition model, a Semantic CV Evaluator, an Automated Cold Outreach tool, and a Real-Time Job Search engine. Prompt engineering with bias awareness mandates an evaluation framework based on skill assessment alone without consideration of any demographic data points during AI evaluations. Evaluation outcomes demonstrate that the PathFinder AI turns the usual disorganized, stressful process of preparation into a well-organized process that enhances candidate preparedness.